Efficient modeling and equalization of uncertain communications channels

P.D. DOÑATE; C.H. MURAVCHIK; J. E. COUSSEAU

Mar del Plata

Congreso; VIII Reunión de Procesamiento de la Información y Control; 1999

RPIC

THis paper presents a methodology for the modeling and decision feedback equalization of uncertain communication channels. The modeling procedure is presented as an algorithm that performs experimental frecuency data to model conversion. A complex exponential base and orthonormal function series are used for the model derivation. The channel model is then used to design a partially adaptive and robust decision feedback equalizer. The fixed fron end or precursor filter is designed as an optimal robust linear deconvolution filter. The feedback filter is designed to optimally cancel residual past symbol interference with a fixed denominator and the numerator is treated as an adaptive FIR filter with a small number of taps. Simulation results show that the equalizer outperforms fixed robust designs and totally adaptive FIR designs. Examples are given showing the fitting characteristics of the modeling approach and the equalizer performance.